Citing this chapter:

Amador, J. A., J. R. Chacón, and S. Laporte, 2003. Climate and climate variability in the

Arenal Basin of Costa Rica. In Climate, Water and Trans-boundary Challenges in the

Americas. Henry Díaz and Barbara Morehouse, Editors. Kluwer Academic Publishers.

Holland, pp. 317-349.

Climate and Climate Variability in the Arenal River Basin of Costa Rica

Jorge A. Amador (*)

Centro de Investigaciones Geofísicas y Escuela de Física,

Universidad de Costa Rica

and

Rafael E. Chacón, and Sadí Laporte

Centro de Servicio de Estudios Básicos de Ingeniería (Area de Hidrología),

Instituto Costarricense de Electricidad y

Centro de Investigaciones Geofísicas, Universidad de Costa Rica

(*) Corresponding author address : Dr. Jorge A. Amador, Centro de Investigaciones

Geofísicas y Escuela de Física, Universidad de Costa Rica. P. O. Box 2060, San José, Costa

Rica.

E-mail: jorge.amador@ucr.ac.cr

2

ABSTRACT

This work examines some of the effects of climate and climate variability in the Arenal River

basin of Costa Rica, the site of the largest hydropower complex in the country. The Arenal

system, which drains part of the north-central portion of Costa Rica covers a total area of

approximately 493 km2; it is mainly managed for electric power generation and produces

nearly a quarter of the total annual electricity in Costa Rica. Monthly, pentad (5-day means),

and daily precipitation data are used to study signals associated with climate and shorter-term

atmospheric disruptions in the basin. Although the study area is relatively small, strong spatial

and temporal contrasts of precipitation patterns are found. A clear distinction in the seasonal

distribution of precipitation is observed over short distances (~30–40 km), between the

northwestern (NW) low lands of the basin compared to the southeastern (SE) sector. The

former region exhibits a bimodal precipitation distribution, with maxima in June and

September–October, and relative minima in July and December–April. The July minimum

suggests a weak mid-summer drought or “veranillo” signal. The latter region has practically

no dry season with the highest precipitation values occurring during the second part of the

calendar year. As determined by principal component analysis of anomalies of monthly

precipitation data, the main disruption of the normal pattern of precipitation appears to be

related to the ENSO signal in the NW region, whereas the SE sector shows a positive

correlation with Caribbean low-level wind changes. Some of the latter changes are associated

with warm or cold ENSO episodes that seem to modulate wind intensities of the low-level jet

over the Caribbean. Precipitation effects in the basin for selected extreme cases, such as that

of hurricane Mitch, and other so-called “temporales” are also analyzed. The importance of

these systems as fundamental components of the basin’s hydrological cycle is well

3

established. ENSO related variability of the regional summer circulation (such as that of the

low-level jet in the western Caribbean), and the appearance of cases of extreme strong trade

winds during the winter circulation are also important forcing mechanisms for precipitation

variability in the basin. In these cases, the interaction between the basin complex topography,

and changes in the flow pattern and intensity seem to be of fundamental importance for

precipitation variance. Some socioeconomic impacts of precipitation variability, as well as a

discussion about the potential use of climate variability information for water management in

the basin are presented.

INTRODUCTION

Costa Rica’s economy has relied in the past mainly on agriculture. During the last

decade, however, some high technology industries have been established increasing the

demand of qualified professionals, and stimulating the adaptation and improvement of higher

education and training programs at universities and technological institutes. Tourism has

become an important source of national income, therefore, building and access to facilities in

natural and ecologically rich environments have expanded substantially. Also, public health

services, and telecommunications are widely spread throughout the country and its more than

three million inhabitants. As expected from this socio-economic situation, there has been an

urgent requirement for providing various forms of energy to users, especially electricity.

Lately, the demand for this type of energy has increased dramatically in Costa Rica, where,

electricity generation is mainly provided by hydropower (75%). Thermal, geothermal, and

wind energy generation (18%, 5%, and 2%, respectively) are the others sources of energy

used to supply electricity to industry and residential users.

4

Climate variability on various time scales, from inter-annual to seasonal, and shorter-

term meteorological phenomena have affected Costa Rica’s economy and jeopardized social

conditions in many different ways during the past. After the severe impacts of El Niño and La

Niña during the last decade, specially those associated with the 1997–98 El Niño event, and

perhaps after the impact of hurricane Mitch in 1998, the general public realized the relevance

and importance of timely meteorological and climate information in the planning,

management and utilization of water resources in the country. A relatively small country,

Costa Rica also possesses a very complex topography (Fig. 1), with distinct annual rainfall

distributions, even on small spatial scales. Large mean annual precipitation totals and suitable

sloping terrain conditions have made hydropower one of the most important natural resources

of the country. The Costa Rica Institute of Electricity (Instituto Costarricense de Electricidad,

ICE), a technical body of the Costa Rica Government created in 1949, has been given the

responsibility of building the infrastructure, and the planning and managing of the water

resources for hydroelectric production. Costa Rica, as has been shown in the last few years, is

very sensitive to water management policies and to changes in water availability. Climate

variability, and in some cases related extreme events, can represent significant challenges to

the long-term well-being of a country that is seeking to develop both economically and

socially. Agriculture is still, of course, a major activity in many rural areas of the country and

a major provider of national income, however, major industries (i.e., electronic industry), and

large commercial factories that also contribute important percentages to the national income,

are located in urban areas and depend on a reliable supply of electricity.

The present contribution deals with the climate setting and with major climate

disruptions affecting the largest hydroelectric system in the country, the Arenal-Corobicí-

5

Sandillal (ARCOSA) complex in the Arenal River basin of Costa Rica. The study is intended

to focus on some regional climate systems affecting the Arenal basin’s own mesoescale

climate, and on how changes in these physical mechanisms are related to disruptions in

precipitation patterns that may have significant effects on the reservoir inflows and water

level, and on electricity production. Despite the importance of the basin for hydropower

generation, relatively few studies have been published regarding meteorological and climatic

conditions in the reservoir and their relationship with regional or global scale phenomena.

Fernández et al. (1986) discussed some of the mesoscale characteristics of the region, and

focused on the reservoir effects on several meteorological parameters around the basin, such

as an increase of locally measured wind velocities. More recently, Amador et al. (2000a,b)

made an extensive study of the disruptions in precipitation patterns due to some selected El

Niño/Southern Oscillation (ENSO), and extreme meteorological events that affected the

basin, especially during the last decade. Some of the major findings of these works will be

discussed below.

There has been general agreement that the skill in ENSO prediction has improved

considerably in the last few years (Hoerling and Kumar, 2000). There is a need, therefore, to

identify and understand major deviations in regional precipitation patterns in order to improve

the application of prediction schemes associated with ENSO events. It is known that El Niño

has an important impact on Costa Rica’s climate, especially along the Pacific slope

(Fernández and Ramírez, 1991; Waylen et al., 1996a). Here, we study precipitation variations

in the Arenal basin for both the 1997–1998 El Niño and the 1998–2000 La Niña events, in the

context of regional and global scale phenomena. The study of any form of variability related

to ENSO or non-ENSO signals requires, however, an adequate description of the “normal”

6

climate conditions of the basin. Therefore, the approach taken here would allow decision

makers to take advantage of any additional information about the above-mentioned

phenomena when planning and operating power generation facilities. In this context, the

seasonal precipitation cycle is examined through analysis of the relative contribution of

different precipitation systems during the course of the year, and the associated rainfall

variability in the basin. The main characteristics of both summer (June-July-August) and

winter (December-January-February) circulations are analyzed and related to relevant features

of the temporal and spatial distribution of sea surface temperature in the adjacent oceans. In

the latter context, the mid-summer drought (Magaña et al., 1999), and the low-level jet over

the Caribbean during the boreal summer (Amador 1998; Amador and Magaña, 1999; Amador

et al., 2000c), constitute two important physical mechanisms associated with the seasonal

cycle leading to some unique characteristics in the distribution of mean monthly precipitation

within the Arenal basin. Finally, in order to show the wide range of climatic variability that

can modulate the spatial and temporal precipitation distribution in the basin, the anomalous

contributions to precipitation of some case studies during the 1997–1998 El Niño and the

1998–2000 La Niña events are presented. A goal of this study is to contribute to a better

understanding of climate and its variability in this important region for hydropower

generation in Costa Rica, with the goal of providing governmental and private institutions

with improved climate information for decision making. By improving the understanding of

the elements that control regional climate and its variability, in relation to both ENSO and

non-ENSO signals, more accurate and tailored climate predictions could be developed to

fulfill some of the needs of this particular socioeconomic sector.

7

In the following, we first present information about the more important hydroelectric

projects in Costa Rica; we then follow with an analysis of general regional climate patterns

and those that pertain specifically to Costa Rica. We then consider the hydroclimatology of

the Arenal River Basin and the large-scale teleconnection patterns that affect the general

climate of the region. The chapter concludes with a summary and discussion.

THE ARENAL BASIN PROJECT

Brief historical account

The Arenal dam and reservoir are located at the southern tip of the Guanacaste

Mountains, in the Guanacaste Province (Fig. 1), some 190 km northwest of San Jose, the

capital city of Costa Rica. The Arenal system is located in a northwest-southeast trending

chain of volcanic cones that is part of Middle America range of active volcanoes. The bedrock

at the dam site, referred to as the “Aguacate formation,” is characterized by a complex of lava

flow rock, tuff, agglomerate, and volcanic breccia, with occasional thin interbeds of sandstone

and shale. Lava flows are known to have blocked the most important local river, the Arenal

River, in the geologic past. The lava flows resulted in deposits, up to 40 meters in thickness,

of fluvio-lacustrine river channel and old flood plain sediments overlying the local bedrock at

all but the highest elevations at the dam site. In recent geological times, most of the dam site

has been covered by a mantle of unconsolidated pyroclastic deposits varying up to 15 meters

in thickness, and consisting of alternating, and often discontinuous layers of volcanic ejecta,

which probably originated from the nearby Arenal Volcano (Wahler and Associates, 1975).

The basin covers an area of approximately 493 km2, about one quarter of the Arenal

Conservation Area. The reservoir is about 17% (84 km2) of the basin’s area, and the current

8

water surface area is about three times that of the original lake. The Arenal River used to

drain the lake towards the Caribbean, as a tributary of the San Carlos and San Juan Rivers, the

latter being part of the border between Costa Rica and Nicaragua. Since the first field studies

a few decades ago, the region showed extraordinary potential for electricity generation,

consequently, the ICE initiated the Arenal project back in 1967. In contrast to the original

Arenal River draining course towards the Caribbean Sea, once the project was completed, the

reservoir was left draining water towards the Gulf of Nicoya in the Pacific Ocean. Other

projected hydroelectric plants, such as Corobicí and Sandillal, were planned to make use of

the water resource in a sort of cascade system. Planning of water use and reservoir system

management included from the beginning, not only the generation of hydropower, but also the

use of water for irrigation projects in the Guanacaste Province, one of the driest regions in

Costa Rica. This area receives tremendous socio-economical benefits from the Arenal-

Tempisque irrigation project, especially, those sectors related to crop activities. The amount

of irrigated land dedicated to agriculture in 1995 was of the order of 16,000 hectares. Positive

changes in productivity after irrigation have been very important for social well being and

farming, since rice and sugar yields have increased from 3 to 8 metric tons per hectare, and

from 8 to 14 metric tons per hectare, respectively.

The Arenal-Corobicí-Sandillal Hydroelectrical Complex

The ARCOSA complex consists of three power generating plants located at different

elevations, namely Arenal, Corobicí and Sandillal, which have a total installed generating

capacity of 362 MW of power. Fig. 2 shows the relative location of the Arenal, Corobicí, and

Sandillal plants with respect to the reservoir. Corobicí alone contributes 174 MW to the

9

system. This capacity has recently been surpassed by the plant at Angostura (177 MW)

located in the Reventazón River on the Caribbean slope of Costa Rica, inaugurated in

December 2000. Arenal plant has a generating capacity of 157 MW. Altogether, the

ARCOSA plants contribute about one quarter of the total annual electrical energy that is

consumed in the country, hence their great importance in the National Electrical System. The

Arenal reservoir has a storage capacity of 2400 million m3

at its maximum operation level of

546 meters above sea level (masl), and has the potential to support electricity demands of

1700 GWh.

Regional Climate Systems

As has been discussed recently by Magaña et al. (1999), the annual distribution of

rainfall contrasts dramatically in the Caribbean side of Central America to that in the Pacific

side. The latter slope exhibits a bimodal distribution, with maxima in June and in September-

October, and a relative minimum during July-August. This reduction in the precipitation is the

so-called mid-summer drought, known locally as the “veranillo” or “canícula”. The existence

of such reduction in the annual distribution of rainfall has been shown to be part of the

seasonal cycle of precipitation over much of Central America (Magaña et al. 1999).

According to these authors, the “veranillo” is related to changes in the intensity of convective

activity over the northeastern Pacific “warm pool”. During the mid-summer drought period,

the trade winds over the Caribbean strengthen, due in part to the dynamical response of the

low-level atmosphere to the magnitude of the convective forcing in the Intertropical

Convergence Zone (ITCZ), which is in turn, associated with the sea surface temperature

(SST) distribution. The starting date, intensity, and duration of the mid-summer drought

10

varies from year to year, and it constitutes an important climate disruption for different sectors

of the economy regarding water availability along much of the Pacific side of Central

America. From December to March, the Pacific slope of Central America enjoys warm and

mostly dry conditions.

The most interesting dynamical feature over the Caribbean Sea during summer is the

low-level jet that develops during June, reaches a maximum in July and weakens in

September (Amador, 1998; Amador and Magaña, 1999). This intense easterly current over the

Caribbean is part of the summer trade wind regime, and constitutes the most important factor

in determining regional climate during this season. The long-term mean (1968–1996) of the

vector wind at 925 hPa (approximately 800 m height) for July, using data from the

NCEP/NCAR Reanalysis Project (Kalnay et al., 1996), is shown in Fig. 3. As can be seen

from the wind field, values in excess of 14 m/s dominate the central Caribbean Sea near 14˚N,

75˚W. The origin of this strong jet has not been fully established, however, its barotropically

unstable nature (Molinari et al., 1997; Amador, 1998; Amador and Magaña, 1999) suggests

that it is through the interaction with transients, such as the easterly waves, that the jet may

obtain part of its kinetic energy. Another possibility may be that this strong flow feeds energy

to the easterly waves, which amplify, increasing their meridional amplitude to the north of its

mean position. Also, from early summer to November, easterly waves and tropical cyclones

constitute major elements among the tropical systems that produce rain over the region. Portig

(1976), and Hastenrath (1991), among others, have documented in some detail the role of

these systems in the annual distribution of precipitation in Central America.

In the Caribbean region, precipitation during the winter months is closely related to

mid-latitude air intrusions (Schultz et al. 1997; 1998) and to low-level cloud systems traveling

11

from the east (Velásquez, 2000). Generally speaking, the winter months along the Caribbean

slope of Central America are wetter and more humid, than conditions prevailing along the

Pacific side, during the same season.

Regarding the beginning and ending of rainy spells, several studies have shown that

fluctuations of the sea surface temperature of the tropical Atlantic and Pacific Oceans are

related to variations in the duration and timing of the rainy season in Central America (e.g.,

Enfield and Alfaro, 1999). Tropical North Atlantic SSTs, for which Alfaro (2000) defined

warm and cold episodes, show the largest influence over the region when compared with that

of other indices, having a positive correlation with rainfall in Central America (Alfaro and

Cid, 1999). In contrast, the Niño3 region, was found to have a lower negative correlation with

the precipitation of the region, influencing only the Pacific slope of Central America.

The Climate of Costa Rica

When dealing with the elements that control climate in Costa Rica, several factors

should be considered. One of these elements, as it could be inferred from the previous section,

is the large influence of the two adjacent oceanic masses on mean precipitation distribution

through the seasonal variation of sea surface temperature and associated circulation patterns

and convective activity. During northern winter, SSTs over adjacent areas of the Caribbean

and the Pacific are relatively uniform, with values usually below 28°C. As a consequence of

this, and the presence of strong vertical trade wind shear, no major convective activity takes

place. Furthermore, the ITCZ is at its southernmost position during the winter months

(Srinivasan and Smith, 1996). Throughout the course of winter, trade winds intensify and

relatively cold air intrusions frequently reach latitudes south of 10°N (Schultz et al., 1997;

12

1998). During summer, the presence of two warm pools dominate the SST distribution, one

over the Caribbean and Gulf of Mexico and the other over the northeastern Pacific, just off the

southern coast of Mexico and west of Central America (Magaña, 1999). In the latter,

organized convective activity is barely observed due mainly to strong subsidence associated

with regional scale circulations, such as those associated with the low-level jet noted earlier.

Other factors, such as the meridional migration of the ITCZ, affects seasonal rainfall

characteristics, especially in southern and central Costa Rica. On the other hand, trade winds,

over the Caribbean side, and south-westerlies on the Pacific side, are important mechanisms

for moisture convergence and rainfall production. The interaction between the main flow and

topographic barriers constitutes a predominant process that contributes to precipitation all

year round.

In countries such as Costa Rica, with very irregular topography, precipitation exhibits

very high temporal and spatial variability. Fig. 4 shows the mean annual distribution of

precipitation over Costa Rica (main map), and the seasonal precipitation patterns at some

selected stations (inserted figures) for different periods. The average annual isohyetal

contours were hand drawn by using 347 stations with precipitation data for the 1970–89

period. Over some mountain basins, such as those of central Costa Rica where precipitation

data is scarce, isohyets were depicted utilizing vertical precipitation profiles from neighbor

basins with enough information that are known to have similar general topographic and

climatic characteristics. Precipitation values estimated by means of this method were

constrained in such a way that they did not exceed the mean runoff of any particular basin.

From the map in Fig. 4, it can be seen that precipitation in the Pacific side varies from about

1400 mm in the northwestern region in the Guanacaste Province, to over 7000 mm in some

13

isolated high elevation areas of Costa Rica. In the Caribbean side, precipitation ranges from

about 1600 mm in the eastern region of the Central Valley and Cartago (to the east of San

José), to 9000 mm in the central Caribbean slope of the northwest-southeast trending

cordillera (see also Fig. 1). Note the relative rainfall maximum (>5000 mm) in the Caribbean

side, near the border between Costa Rica and Nicaragua. This maximum, in fact, extends well

along the Nicaragua coast to the north reaching southeastern Honduras (Amador and Magaña,

1999). At first sight, it seems that this maximum may be due to orographic forcing of the trade

wind flow, however, as can be observed from Fig. 1, the terrain over that area is almost flat.

The physical processes responsible for this rainfall maximum have yet to be fully explained;

however, as suggested by Amador (1998), at least, a major factor in summer rainfall could be

due to the low-level convergence associated with the jet exit over the western Caribbean (Fig.

3).

To show that this trade wind current forms part of the basin’s fundamental climate

features, Fig. 5 is presented. Using data from the Pan American Climate Studies Sounding

Network (PACS-SONET) pilot balloon project at Liberia, located some 50 km west of the

reservoir. Fig. 5 shows, for July 1997, the vertical structure of the jet downstream of the

basin’s region. This intense current attains its maximum values of about 10 m/s, just at the

level of the basin’s main topographical features (1500 to 2000 m).

Average precipitation value for Costa Rica is 3300 mm using data for the period

1970–89. Fig. 4 also shows, as inserted diagrams, the distribution of mean monthly

precipitation for some selected regions of Costa Rica. Different types of annual rainfall

distributions can be observed. In the north Pacific region (Bagaces), a pronounced dry period

extending from December to April dominates the Northern Hemisphere (NH) winter season,

14

with a well defined rainy spell from May to November. A secondary minimum associated

with the “veranillo”, can be seen clearly during July-August. In the south Pacific region

(Potrero Grande), the dry period from December to March is less marked than that in the

North Pacific sector. The wet period, on the contrary, is more pronounced in southern Costa

Rica, but the “veranillo” signal can still be observed.

The contrast between Pacific and Caribbean rainfall distributions is noteworthy. At the

eastern Caribbean region (Bataán), rainfall shows two minima, one in March and the other

one in September–October. The maximum rainfall occurs in July, probably associated with

the Caribbean low-level jet that develops during this period. It may appear that orography

plays an important role in the occurrence of this maximum, however, Bataán is located in an

almost flat terrain, some 50 km to the east of the mountain range. A secondary maximum is

observed in December as a result of the intensification of the trade winds, and the southward

displacement of air masses, during the Northern Hemisphere winter. These relatively cold

northerly winds that appear especially during December–February are often referred locally,

as “los nortes” (Portig, 1976). Finally, the northern region of Costa Rica (Ciudad Quesada)

shows a distribution with a minimum in March – April. The maximum rainfall tends to occur

in July, although, during most of the wet season precipitation is relatively large.

HYDROCLIMATE OF THE ARENAL RIVER BASIN

The average rainfall pattern over the Arenal basin is presented in Fig. 6. The

distribution was estimated using 25 gauging stations (6 conventional, 19 automatic) located

within the basin and adjacent regions for the 1970–1999 period. To ensure a proper rainfall

analysis, the station precipitation records were verified by means of the double accumulation

15

curve method, and those sites containing errors of a systematic type were corrected. From Fig.

6, it can be seen that annual rainfall varies from a maximum of 5000 mm in the southeastern

(SE) region to 2000 mm in the northwestern (NW) lower side of the basin. A secondary

maximum of 4000 mm is located in the higher lands of the northwestern region. The above

mentioned rainfall maxima are due partly to the interaction of the trade wind regime with the

mountain ranges having a southeast to northwest orientation. Rainfall tends to diminish

towards the western basin sectors, comprising the driest region of Costa Rica in the

Guanacaste Province with just over 1400 mm per year (see Fig. 4).

Although the study area is relatively small, strong spatial and temporal contrasts of

precipitation patterns are found (Fig. 7). A clear distinction in the seasonal distribution of

precipitation is observed over short distances (~30–40 km), between the NW lower part of the

basin (represented by Naranjos Agrios rainfall data, hereafter NA) as compared to the SE part

(represented by Caño Negro data, hereafter CN). As observed in Fig. 7a, using pentad data,

the NW region exhibits a bimodal precipitation distribution with maxima around May and

September-October, and a relative minimum in late June to late July. This minimum suggests

a weak mid-summer drought or “veranillo” signal. The SE region, presented in Fig. 7b, has

only a relatively short dry season with the highest precipitation occurring during the second

part of the calendar year. In this distribution, a rainfall peak is noticeable during the summer

months, which indicates the importance of the interaction of the trade winds, associated with

the Caribbean low-level jet, with the basin topographic features. The analysis also suggests

that changes in the intensity of the low-level jet could be an important mechanism for

precipitation variability during the summer months, which in turn, may affect inflows to the

reservoir.

16

Additional hidroclimatological information for the Arenal Basin is presented in Fig. 8.

In Fig. 8a, monthly values of extreme air temperatures for two transects are shown. Nueva

Tronadora (NT) is located in the NW lower lands, and Presa Sangregado (PS) is situated in

the SE part of the basin. As discussed earlier based on Fig. 7, the above two regions show the

largest contrast in monthly precipitation distribution within the basin. As was expected,

changes in radiation and cloudiness are responsible for extreme surface temperature behavior.

Note that absolute maximum temperature (Tmax) corresponds well with reduced periods of

precipitation, from March to April approximately, in both regions of the basin (see Fig. 7). As

rainfall increases during April–May, as a consequence of the onset of the rainy season and the

ITCZ northward migration, maximum temperature drops in both sites to a relative minimum

in July, suggesting a weak or masked “veranillo” signal. Since the presence of the mid-

summer drought (Fig. 7a) implies a relative reduction of precipitation and cloudiness, Tmax

should show a relative maximum during the mid-summer drought, as opposed to what is

observed in the NW sector of the basin. A plausible explanation for above-noted behavior of

Tmax could be the presence of strong winds associated with the low- level jet during the

summer season which may act to lower temperature somewhat. Minimum temperatures

(Tmin) are consistent with the development of the dry and wet seasons, but they do not show

any significant changes that could be associated with the “veranillo”, as was the case shown

by Magaña et al. (1999). During July–August an increase in Tmin is indicated for station PS

that could be related to greater atmospheric moisture content associated with an increase in

cloudiness and rainfall, due to the interaction of the strong summer winds with topography in

the SE sector of the basin.

17

As discussed earlier, the increase in mean precipitation during the summer months

over the eastern part of the basin appears to be associated with the development of the low-

level jet in the Caribbean by means of the interaction of the mean wind and the eastern most

topographical features of the basin. In Fig. 8b, mean streamflow for two stations (El Cairo,

EC, for 1975–2000), and Nueva Arenal, AN for 1978–2000) is presented. The presence of a

maximum in this parameter during July–August, corresponding to the period of maximum

wind speed associated with the low-level jet over the Caribbean can be seen clearly in Fig. 8b.

Two minima are found in this parameter at both stations, one in April and the other one in

September, the former corresponding to the dry spell along the Pacific slope of Costa Rica

and the latter occurring just after the drop of the trade wind intensity in September (Amador,

1998). After June-July, the intense interaction of the strong low-level trade winds associated

with the low-level easterly jet with topography, produces an important contribution to mean

precipitation in the SE sector of the basin, and as a consequence of that, there is a substantial

seasonal increase in the reservoir water level. Both, the decrease in the streamflow, and the

reduction in the intensity of the low-level jet, suggest that an index associated with this

intense low-level flow over the Caribbean should be included in any prediction scheme of

precipitation variability in the basin.

PHYSICAL MECHANISMS OF CLIMATE VARIABILITY

IN THE ARENAL RIVER BASIN

Empirical orthogonal function (EOF) analysis of monthly rainfall anomalies was

performed for two different subsets of station data in the basin. The subsets of stations were

subjectively chosen following the two main seasonal precipitation distributions observed in

18

the basin (see Fig. 7). In the NW lower terrain and relatively drier region, Naranjos Agrios

(10° 32'N, 84° 59'W), Nueva Tronadora (10° 30'N, 84° 55'W), Toma de Arenal (10° 30'N, 84°

59'W), La Tejona (10° 31'N, 84° 59'W, Dos Bocas (10° 33'N, 84° 55'W), and Aguacate (10°

33'N, 84° 57'W) were used for the EOF analysis during the period 1979–1998. In the SE,

Presa Sangregado (10° 29'N, 84° 46'W), Pueblo Nuevo (10° 26’N, 84° 47’W), Pastor (10°

25’N, 84° 45’W), Pajuila (10° 30’N, 84° 47’W), Jilguero (10° 27'N, 84° 43'W), and Caño

Negro (10° 24'N, 84° 46'W) were utilized also during the above mentioned period (use Fig. 6

to approximately locate the stations used). The first principal component of precipitation

anomalies for NW and SW regions explains 72% and 83% of this variable, respectively.

Cross-correlation analysis of the dominant mode for both regions was carried out with

corresponding anomalies from several global and regional indices, such as, the SST index for

the Niño3-4 region, North Atlantic Oscillation (NAO), Southern Oscillation Index (SOI), and

925 and 700 hPa mean wind speeds. The latter parameter was averaged over the area 10°–

20°N, 60°–80°W, as a simple index of the mean intensity of the low-level trade wind regime.

Broadly speaking, the correlations are relatively small for both regions, indicating the

presence of other dominant mechanisms for rainfall variability. The following aspects can,

however, be identified from EOF analysis. From the estimates of the correlations for lags up

to six months that are significant at a 99% confidence level, rainfall changes in the NW drier

region have a maximum correlation coefficient of –0.23 for a lag of two months with El

Niño3-4 index. In other words, changes in SST over the central Pacific associated with warm

(cold) ENSO events lead for nearly two months to abnormally dry (wet) conditions in the

Pacific side of the basin. Regarding the SE region of the basin (Caribbean side), precipitation

anomalies do not seem to be significantly correlated to SOI or SST’s in the Pacific. Waylen et

19

al. (1996b) in their study of time and space variability of annual precipitation in Costa Rica in

relation to SOI, noted that there is a marked difference in response in those areas draining

towards the Pacific and those towards the Caribbean, which in a broad sense, agrees well with

the results of this study. George et al. (1998) also found that annual discharge for rivers

within the Pacific watershed were positively associated with SOI values, whereas in the

Caribbean, discharges showed less clear and coherent patterns of associations. They also

suggested that this difference in response could be related to the elevation of the river basins.

The complexity of the response in relation to topography has been pointed out by Waylen et

al. (1996b), in reference to the differences in dominant precipitation generating mechanisms

within the basins. For the SE region, however, 925 hPa wind speed changes averaged over the

area 10°–20°N, 60°–80°W, are positively correlated (0.24) at a 99% confidence level, with

disruptions of rainfall from the normal pattern at zero lag. Stronger (weaker) than normal

winds appear to be responsible for wetter (dryer) than average conditions over the Caribbean

side of the basin. In order to gain some insight into the nature of the prevailing precipitation

originating mechanisms, the seasonality of the above results is analyzed below using a finer

temporal precipitation resolution for the phases of El Niño 1997–98 (approximately, from

March 1997 to June 1998) and La Niña 1998–2000 (from July 1998 to July 2000).

Fig. 9a,and Fig. 9b, present the normalized pentad precipitation anomalies for NA and

CN, respectively, for El Niño 1997–1998, and Fig. 9c, and Fig. 9d for NA and CN,

respectively, for La Niña 98–2000. Pentad precipitation anomalies were estimated as

departures from the mean values for the period of analysis and normalized by the

corresponding standard deviation. For both stations, during El Niño 1997–98, relatively long

sub-periods of negative anomalies can be clearly identified. From the annual perspective, the

20

impact of this event was to produce weak below average conditions as a whole, implying a

deficit on water availability for electricity generation. These results are consistent with those

of Amador et al. (2000a) for El Niño events of 1992–93 and 1994–95, although the physical

mechanisms responsible for this kind of response have yet to be proposed. Previous results

from principal component analysis discussed earlier suggest the importance of identifying the

dominant regional climate variability modes to explain changes in precipitation within the

basin.

We have noted that precipitation anomalies are associated with changes in trade wind

intensity and in the low-level jet, which are also associated with ENSO episodes. Figure 10

shows the composite anomaly of the 925 hPa wind vector associated with the low-level jet

over the Caribbean during several El Niño, (Fig. 10a), and La Niña, (Fig. 10b) events, for the

summer months (June to August). Composite anomalies of the wind vector were estimated

using at least 9 episodes for each of the two ENSO phases. El Niño and La Niña events were

defined following Kiladis and Diaz (1989), and using a procedure similar to that proposed by

Trenberth (1997). El Niño (La Niña) summers are characterized by stronger (weaker) than

normal winds over the central Caribbean associated with the low-level jet core region. From

Fig. 9b, a marked period of positive precipitation anomalies is observed, at CN during July-

August 1997, which is consistent with the idea discussed earlier, that changes in wind

intensities are associated with variations in the low-level jet, that in turn are related to

precipitation anomalies over the basin through a wind–topography interaction mechanism.

Note that in NA (Fig. 9a), July–August 1997 exhibit mainly negative precipitation anomalies,

as expected from a strong descending wind flow over the NW sector of the basin. El Niño

event during summer implies a stronger low-level jet, that are reflected in positive

21

precipitation anomalies in the Arenal Basin, especially in the SE sector, which dominates the

contribution of precipitation to the reservoir water level this time of year. Kiladis and Diaz

(1989) detected a relatively weak trend toward drier than average conditions in the Caribbean

basin from July to October of a canonical warm ENSO phase. Their result is consistent with

the one found here, at least for the western Caribbean, since conditions are unfavorable for

convection due to the strengthening of the trades during El Niño summer months, to increased

vertical wind shears (Amador et al. 2000c), and to related negative SST anomalies due in part

to enhanced evaporation. Once the trade winds weaken in September-October, the SSTs start

to recover slowly towards the end of the rainy season, favoring again convective activity

before the start of the winter months. Negative anomalies during El Niño winter of 1998 are

discussed below in conjunction with results shown in Fig. 11.

Positive anomalies for other periods at NA and CN during the 1997–1998 El Niño

correspond to the so-called “temporales” and mid-latitude air intrusions. Some of these cases

are discussed later, in more detail, on this chapter. Fig. 9c and 9d, show for the NA and CN

regions, respectively, the pentad precipitation anomalies during the most recent La Niña event

(1998–2000). Contrary to what is observed in Fig. 9a and 9b, the 1998–2000 cold phase of

ENSO presents both positive and negative anomalies with no apparent relationship with

global or regional scale systems. A more detailed analysis, however, reveals relatively dryer

than normal conditions at NA and CN during the summer months (July–August 1998, and

July–August 1999). Fig. 10b implies that the low-level jet during summer of a cold ENSO

phase is associated with weaker than normal wind velocities, which is consistent with the

negative summer anomalies found at both stations in Fig.9c and 9d, under the assumption of

weaker flow-topography interactions. Also, one could argue that weaker (stronger) trades

22

associated with the low-level jet entail decreased (increased) influx of atmospheric moisture

content into Costa Rica coming from the Caribbean, and as a consequence into the Arenal

Basin. Changes in the strength of the low-level jet lead to favorable (unfavorable) conditions

for the generation of rainfall. A similar mechanism that partially explains hydrological

anomalies in some areas of the Pacific slope of Colombia is also associated with a low-level

westerly jet that penetrates inland, especially from August to November, with maximum

winds of about 6 m/s in October (Poveda et al. 1998; Poveda and Mesa, 2000).

El Niño and La Niña composites for the 925 hPa vector wind anomalies, for NH

winter are shown in Fig. 11a and 11b, respectively. The warm (cold) ENSO phase shows

weaker (stronger) than normal wind velocities over the Caribbean region that is reflected in

less (more) precipitation at both stations (Fig. 9a,b). The exception to the above statement is

December 1998 to February 1999 for CN (Fig. 9d), which displayed near normal conditions.

Noteworthy are the positive precipitation anomalies of the winter months of 1999–2000. A

discussion of one of the extreme cases studied for the winter months of 1999–2000 follows.

Some case studies of short-term meteorological phenomena that hit the Arenal Basin

are presented in Fig. 12. The aim is to show their relative importance to precipitation

variability, and to identify these types of systems as fundamental components of the basin’s

hydrologic budget. Figure 12a shows the contribution to local precipitation of several extreme

cases during the 1997–98 El Niño event, namely, the “temporales” of 1–6 August 1997, and

of 9–13 March 1998. The name “temporales” is the term used locally for a period of weak to

moderate nearly continuous rain covering several days affecting a relatively large region.

Hastenrath (1991) provides a definition, of the “temporales” for the Pacific Region of Central

America, close to the one used here, however, a condition that is included in his concept is

23

that winds are weak, while in some cases, as it is shown below, winds could be intense and

long lasting. The frequency of these events shows a great deal of inter-annual and intra-

seasonal variability, and their relationship to ENSO or to other large-scale climatic signals is

still unclear. As discussed by Velásquez (2000), these perturbations not only originate in the

Pacific as disturbances associated with the ITCZ, also discussed in Hastenrath (1991), but are

also related to westward traveling low-level cloud systems over the Caribbean, not always

associated with mid-latitude air intrusions. That is the case of the August 1997 episode and

the beginning of the January 2000 event (the latter shown in Fig. 12b). It can be noted in Fig.

12a that CN experiences a dramatic increase in rainfall that surpasses 300 mm in about 5 days

for the August 1997 temporal. Other stations (not shown) also measured important amounts of

precipitation related to this temporal. Note from Fig. 12a that NA in the NW low lands of the

basin does not show an important response to this Caribbean temporal. Another case that

confirms this difference in the response of the basin to Caribbean temporales is included in

Fig. 12a for March 1998. The January 2000 case is initially characterized by a sudden wind

intensification over the Caribbean from 5 to 9 January 2000, as illustrated in Fig. 13a. To

show that this wind change is not initially related to cold air intrusions from mid-latitudes,

Fig. 13b is shown. This case corresponds to the period 11–20 January 2000 that affected

almost the whole basin (Fig. 12b). Due to the anomalous rainfall of late 1999 in the SE sector

(see Fig. 9d), and to the temporal of January 2000, the reservoir level rose to an

unprecedented level of nearly 548 m. A safe level of operation, according to ICE is 546 m,

and therefore emergency measures had to be taken regarding the water use and management

for electricity generation. The photograph presented in Fig. 14 illustrates the overall impact of

the anomalous precipitation on the water level of the lake. Note that this temporal, contributed

24

more than three times the precipitation attributed to Hurricane Mitch at CN during 26–30

October 1998 (Fig. 13b), which occurred during a cold ENSO event. After this extreme event

hit Costa Rica’s northern region in January 2000, overall bean crop losses due to flooding and

severe meteorological conditions were reported to be of the order of US$3 million.

The cases presented before provide evidence that the topographical features of the

basin play a crucial role as a forcing mechanism for generation of rainfall. Changes in low-

level trade wind intensity are closely related to disruptions in precipitation by means of a

flow-topography interaction process. The local dependence of precipitation amount upon

elevation in areas of high relief in Costa Rica has been well established (Chacón and

Fernández, 1985, Fernández et al. 1996), however, the relationship of wind intensity to

precipitation amount still requires further investigation.

SUMMARY DISCUSSION

The use of pentad precipitation data and the availability of a relatively dense station network

to update aspects of the climatology of the Arenal River Basin, clearly has helped to improve

our understanding of the seasonal distribution of precipitation for water management

purposes. It was also possible to identify the areas with strong spatial and temporal contrasts

in precipitation patterns for water use purposes. A relative minimum in the annual cycle of

precipitation, known as the mid-summer drought or “veranillo”, weakly affects the most

western portion of the reservoir during July-August. Intense trade winds, associated with the

development of a low-level jet over the Caribbean during the summer months, have an

opposite effect to that of the veranillo. Increases in the wind flow results in an increase of

25

precipitation in the eastern most region of the reservoir, due mainly to the interaction of the

wind flow with the basin’s topography.

The 1997–98 El Niño and 1998–2000 La Niña episodes have been closely studied in

relation to their influence on the basin precipitation distribution. Based on previous results by

Amador et al. (2000a,b) and those of the present work, El Niño events are shown generally to

be associated with a decrease in reservoir level, although some other factors also affect this

parameter. La Niña events generally increase the precipitation in the basin, especially that of

the northern and western parts, by means of disturbances originating in the Caribbean side,

such as the so-called temporales and hurricanes.

The first order climate disruption from the normal precipitation pattern in the Arenal

basin, as determined by principal component analysis of anomalies of monthly precipitation

for two regions (NW and SE), appears to be weakly related to ENSO episodes and to a greater

degree to low-level Caribbean wind anomalies, respectively. Although these low-level wind

changes seem to be modulated partly by ENSO, results bring out the importance of other non-

ENSO signals, such as variations in the low-level jet, for understanding climate disruptions in

the region.

Another factor that contributes to inter-annual precipitation variability in the basin is

one related to the frequency of cyclone formation in the Caribbean basin. It is generally

accepted that during El Niño episodes, fewer tropical cyclones tend to form in the Caribbean

region. Amador et al. (2000c) found, on the one hand, that the existence of anomalously warm

SST's in the eastern Pacific results in an enhanced low-level jet, and therefore, in stronger

than usual wind shear during summer. On the other hand, during La Niña years, the low-level

jet weakens and the vertical wind shear decreases. Partly as a result of such changes in the

26

wind shear environment, the number of hurricanes varies from one year to another. The

fluctuations in the intensity of the low-level jet are also reflected in the SST anomaly field

over the western Caribbean Sea, north of the Venezuelan coast. An intense (weak) jet results

in negative (positive) SST anomalies over the western and central Caribbean, due to strong

(weak) Ekman transport and upwelling. In this way, the low-level jet may also be playing a

role relating the SST anomalies in the eastern Pacific during El Niño or La Niña events, to the

SST anomalies over some regions of the Caribbean.

The importance of identifying precipitation variability on inter-annual time scales,

other than ENSO, associated with regional climatic features, such as the “veranillo” and the

low-level jet, has led us to identify basic steps in the development of practical rainfall

forecasting schemes for the Arenal basin. The problem of climate variability in Costa Rica

and other regions of Central America constitutes a challenge, since the region is surrounded

by warm ocean pools in which convective activity is intense, sometimes as a result of the

interaction of strong trade winds associated with regional climate systems in complex

topography. Most forecasting schemes aimed at determining the spatial patterns of

precipitation anomalies are based on statistical methods or on analogues, using the El Niño

signal as primary input. As has been shown here, not all the anomalous climate patters exhibit

the same spatial characteristics, or are related to the ENSO signal directly. Even more,

differences in large-scale climatic patterns associated with different ENSO events sometimes

makes the utilization of analogues or multiple regression models based only on tropical

Pacific SST information of very limited use. Therefore, the use of more physically based

prognostic tools, such as that provided by numerical models with a multi-parameter output,

27

should in the future be incorporated as a more realistic alternative to the traditional

predictions schemes.

Acknowledgements

This work was partially funded by the University of Costa Rica (grants VI-805-94-204 / 98-

506 / 112-99-305), and the Inter-American Institute for Global Change Research (IAI) grants

ISP3-030 and CRN-073. Additional support from National Oceanic and Atmospheric

Administration (Office for Global Programs) and Comité Regional de Recursos Hidráulicos

through a project for regional modeling is also acknowledged. Thanks are due to Dr. Victor

Magaña, National Autonomous University of Mexico, for many fruitful discussions on

regional climate systems. In addition, we would like to thank Dr. E. Alfaro and Dr. F. J.

Soley, University of Costa Rica, for helpful talks over practical aspects of EOF. The authors

are also indebted to E. Rivera, R. Velásquez, I. Mora, R. Madrigal, Z. Umaña, and other staff

members of CIGEFI, for their assistance during the course of this work. Reviewers provided

many valuable comments to improve the quality of the manuscript. ICE and Instituto

Meteorológico Nacional kindly provided Arenal River Basin, and Costa Rica climate data.

Use of NCEP/NCAR public domain data facilitated part of this research.

List of Abbreviations

AN: Arenal Nuevo Hydrometeorological Station

ARCOSA: Hydroelectric Complex Arenal-Corobici-Sandillal

CN: Caño Negro Hydrometeorological Station

EC: El Cairo Hydrometeorological Station

IAI: Inter-American Institute for Global Change Research

28

IGN: Instituto Geográfico Nacional (National Geographic Institute)

ICE: Instituto Costarricense de Electricidad (Costa Rica Institute of Electricity)

NA: Naranjos Agrios Hydrometeorological Station

NCEP: National Center for Environmental Prediction

NCAR: National Center for Atmospheric Research

NT: Nuevo Tronadora Hydrometeorological Station

PACS-SONET: Panamerican Climate Studies Sounding Network

PS: Presa Sangregado Hydrometeorological Station

REFERENCES

Alfaro, E. J., y L. Cid, 1999: Ajuste de un modelo VARMA para los campos de anomalías de

precipitación en Centroamérica y los índices de los océanos Pacífico y Atlántico

Tropical. Atmósfera, 12, 205–222.

Alfaro, E., 2000: Eventos Cálidos y Fríos en el Atlántico Tropical Norte. Atmósfera, 13, 109–

119.

Amador, J. A., 1998: A climatic feature of the tropical Americas: The trade wind easterly jet.

Tópicos Met. y Ocean., 5(2), 91-102.

Amador, J. A., and V. O. Magaña, 1999: Dynamics of the low level jet over the Caribbean

sea. Preprints 23th

Conference on Tropical Meteorology, 10–15 January 1999, Dallas,

Texas, Amer. Meteor. Soc., 868–869.

Amador, J. A., S. Laporte, and R. E. Chacón, 2000a: Cuenca del Río Arenal: Análisis de los

eventos El Niño de los años 1992-93, 1994-95 y 1997-98. Tópicos Met. y Ocean., 7(1), 1-

21.

Amador, J. A., R. E. Chacón, and S. Laporte, 2000b: Cuenca del Río Arenal: Análisis de los

eventos La Niña de los años 1988-89 y 1996. Tópicos Met. y Ocean, 7(1), 22-42.

29

Amador, J. A., V. O. Magaña, and J. B. Pérez, 2000c: The Low level jet and convective

activity in the Caribbean. Preprints 24th

Conference in Hurricanes and Tropical

Meteorology, 29 May–2 June 2000, Fort Lauderdale, Fla., Amer. Meteor. Soc., 114–115.

Chacón, R. E., and W. Fernández, 1985: Temporal and spatial rainfall variability in the

mountainous region of the Reventazón River basin, Costa Rica. J. Climatol., 5, 175–188.

Enfield, D., and E. J. Alfaro, 1999: The dependence of Caribbean rainfall on the interaction of

the tropical Atlantic and Pacific Oceans. J Climate 12, 2093–2103.

Fernández, W., R. E. Chacón, and J. W. Melgarejo, 1986: Modifications of air flow due to the

formation of a reservoir. J. Clim. Appl. Meteorol., 25, 982–988.

Fernández, W., and P. Ramírez, 1991: El Niño, la Oscilación del sur y sus efectos en Costa

Rica : Una revisión. Tecnol. Marcha, 11, 3–11.

Fernández, W., R. E. Chacón, and J. W. Melgarejo, 1996. On the rainfall distribution with

altitude over Costa Rica. Revista Geofísica, 44, 57–72.

George, R., P. Waylen, and S. Laporte, 1998: Interannual variability of annual streamflow and

the Southern Oscillation in Costa Rica. Hydrol. Sci ., 43, 409–424.

Hastenrath, S., 1991: Climate Dynamics of the Tropics. Kluwer Academic Publishers, 488 pp.

Hoerling, M. P. and A. Kumar, 2000: Understanding and Predicting Extratropical

Teleconnections Related to ENSO. In: H. F. Diaz and V. Markgraf (eds.), El Niño and

the Southern Oscillation, Multiscale Variability and Global and Regional Impacts.

Cambridge University Press, 57–88.

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR Reanalysis 40-year Project. Bull. Amer.

Meteor. Soc. 77, 437–471.

30

Kiladis, G.N., and H.F. Díaz, 1989: Global climatic anomalies associated with extremes in the

Southern Oscillation. J. Climate, 2, 1069–1090.

Magaña, V.,1999: Los Impactos de El Niño en México. Universidad Nacional Autónoma de

México y Dirección General de Protección Civil, Secretaría de Gobernación, México,

229 pp.

Magaña, V. O., J.A. Amador, and S. Medina, 1999: The mid-summer drought over Mexico

and Central. J. Climate, 12, 1577–1588.

Molinari, J., D. Knight, M. Dickinson, D. Vollaro, and S. Skubis, 1997: Potencial vorticy,

easterly waves and eastern Pacific tropical cyclogenesis. Mon. Wea. Rev., 125, 2699–

2708.

Portig, W.H., 1976: The climate of Central America. In World survey of Climatology,

Volume 12, Climate of Central and South America, Ed. W. Schweidtfeger. Elsevier

Publishing Company, 405–478.

Poveda, G., M. Gil, and N. Quiceno, 1998: El Ciclo Anual de la Hidrología de Colombia en

Relación con el ENSO y la NAO. Bull. Inst. Fr. d’Etudes Andines, 27(3), 721–731.

Poveda, G. and O. J. Mesa, 2000: On the existence of Lloró (the rainiest locality on Earth):

Enhanced ocean-atmosphere-land interaction by a low-level jet. Geophysical Research

Letters, 27(11), 1675–1678.

Schultz, D., W. E. Bracken, L. Bosart, G. Hakim, M. Bedrick, M. Dickinson, and K. Tyle,

1997: The 1993 Superstorm cold surge: frontal structure, gap flow, and tropical impact.

Mon. Wea. Rev., 125, 5–39.

Schultz, D., W.E. Bracken, and L. Bosart, 1998: Planetary and synoptic-scale signatures

associated with Central American cold surges. Mon. Wea. Rev., 126, 5–27.

31

Srinivasan, J., and G. Smith, 1996: Meridional migration of Tropical Convergence Zones. J.

Climate, 9, 1189–1202.

Tremberth, K.E., 1997: The definition of El Niño. Bull. Amer. Met. Soc., 78, 2771–2777.

Velásquez, R. C., 2000: Mecanismos físicos de variabilidad climática y eventos extremos en

Venezuela. Tesis de Licenciatura en Meteorología, Departamento de Física Atmosférica,

Oceánica y Planetaria, Escuela de Física, Universidad de Costa Rica, 118 pp.

Wahler, W. A., and Associates, 1975: Design report for Arenal Dam. Volumes 1, 2, Project

Summary, California, USA, 154 pp.

Waylen, P., C. Caviedes, and M. Quesada, 1996a: Interannual variability of monthly

precipitation in Costa Rica. J. Climate, 9, 2606–2613.

Waylen, P., M. Quesada, and C. Caviedes, 1996b: Temporal and spatial variability of annual

precipitation in Costa Rica and the Southern Oscillation. Int. J. Climatol., 16, 173–193.

32

Figure Legends

Fig.1. Topographic map of Costa Rica showing the approximate location of the Arenal Basin with

respect to the town of Liberia in the Guanacaste Province, and San José, the capital city of the country.

Altitude is given in meters above sea level (masl).

33

Fig. 2. Relative location of the Arenal, Corobici, and Sandillal hydroelectric plants with respect to the

Arenal reservoir.

34

Fig. 3. Long term 925 hPa mean wind vector (m/s) showing the location of the low level jet over the

Caribbean during July using NCEP/NCAR Reanalysis data for the period 1968-1996.

35

Fig. 4. Mean annual distribution of precipitation over Costa Rica (main map) for the period 1970-

1989, and seasonal rainfall distribution at some selected stations (inserted figures) for shown periods.

Precipitation contours are in mm per year and intervals are shown in gray scale.

36

Fig. 5. Vertical structure of the mean zonal component (m/s) of the low level jet, using PACS-SONET

pilot balloon data at Liberia, Costa Rica, for July 1997.

37

Fig. 6. Mean annual distribution of precipitation in the Arenal Basin for the period 1970-1999

showing, approximately, the hydrometeorological station network used in this study (black dots). The

stations, for which data is explicitly discussed, are NA (Naranjos Agrios), NT (Nueva Tronadora), CN

(Caño Negro), EC (El Cairo), PS (Presa Sangregado), and AN (Nuevo Arenal). Green solid line shows

approximately the boundaries of the Arenal reservoir. Isohyets are contoured at 400 mm intervals from

2000 to 4000 mm, and at 1000 mm intervals above 4000 mm.

38

Fig. 7. Mean pentad precipitation distribution for a) Naranjos Agrios for the period 1974-

2000, and b) Caño Negro for the period 1966-2000. Precipitation units are mm per pentad.

39

0

2

4

6

8

10

12

14

J F M A M J J A S O N D

months

EC streamflow (m3/s)

0

0.5

1

1.5

2

2.5

AN streamflow (m3/s)

EC mean streamflow AN mean streamflow

Fig. 8. Mean monthly distribution of a) maximum and minimum temperatures (Tmax and Tmin,

respectively, both in °C) at Nueva Tronadora (NT) for 1978-1999, and Presa Sangregado (PS) for

1983-1999, and b) streamflow in m3/s at El Cairo (EC) for 1975-2000, and Nuevo Arenal (NA) for

1978-2000.

40

Fig. 9. Normalized precipitation anomalies (mm) by pentad for a) El Niño 1997-1998 at Naranjos

Agrios, b) as in a) but for Caño Negro, c) La Niña 1998-2000 at Naranjos Agrios, and d) as in c) but

for Caño Negro.

41

a)

b)

Fig. 10. Composite anomaly of the 925 hPa wind vector (m/s) for a) El Niño summers (June to

August), and b) La Niña summers, using NCEP/NCAR Reanalysis data. El Niño and La Niña events

follow basically the definition provided by Kiladis and Diaz (1989) and the procedure proposed by

Trenberth (1997).

42

a)

b)

Fig. 11. As in Fig. 10 but for a) El Niño winters (December to February), and b) La Niña winters.

43

Fig. 12. Total precipitation contribution (mm) of selected “temporales” and extreme meteorological

events at Caño Negro (CN) and Naranjos Agrios (NA) during a) El Niño 1997-1998, and b) La Niña

1998-2000.

44

a)

b)

Fig. 13. Average values at 925 hPa over the area 10-15°N,75-80°W from 5-25 January 2000 for a) air

temperature (°K), and b) wind speed (m/s).

45

Fig. 14. Photograph of the Arenal Lake during the “temporal” of 11-20 January 2000. Note the

location of an approximately 6 m high tree seen in the lake, which used to be along the lakeshore

before the unprecedented water level rise due to the temporal and La Niña rainfall anomalies of late

1999 and early 2000. Photo taken by J.A. Amador on January 15, 2000.